Life, but not as we know it

Thanks to the new science of synthetic biology, it will soon be possible to create living cells in a laboratory. This could bring big benefits—from medicine to combating global warming—but potential dangers too. I went to Greenland to find out more
July 31, 2007

I have just spent a week 120 miles above the Arctic circle listening to a hand-picked group of 17 scientists discuss how to make synthetic life forms. Like it or not, this is going to happen, possibly in the next few years. Some will find that shocking, even blasphemous. To others it will seem a tremendous opportunity, scientifically and economically. Some hope it will help to solve urgent global problems. In any event, there is clearly some explaining to be done.

We are not the first to imagine making life anew to our own design. In Francis Bacon's New Atlantis (1627), the scientist-priests who rule the technological utopia of Bensalem on a Pacific island reveal how life has become clay in their hands:

We make by art, trees and flowers to come earlier or later than their seasons: and to come up and bear more speedily than by their natural course they do. We make… their fruit greater and sweeter and of differing taste, smell, colour and figure… We have also parks and enclosures of all sorts of beasts and birds… By art likewise we make them greater or taller than their kind is, and contrariwise dwarf them and stay their growth. We make them more fruitful and bearing than their kind is, and contrariwise barren and not generative…

It is barely possible to read this today without revulsion, although Bacon deemed these dreams to be a good thing. His vision of a research programme that might lead to such "art" served as a template for the formation of the Royal Society in London in the mid-17th century.

article body image

Nearly 400 years later, Bensalem's mastery of life is nigh. Of course, selective breeding can already make beasts "greater or taller," and genetic engineering has produced plants that grow faster, or out of season, or altered in taste and colour. But this is mere tinkering, often slow or unpredictable. The new life forms now being reported are decidedly more Baconian in their artificiality: bacteria that pulse on and off with light, or that act as photographic emulsion, or that "count" the number of times they divide, or that enact computer logic.

These are the products of synthetic biology, an emerging science that breaks down life as we know it into its component parts, and then reassembles them into something new. The size of these biological parts is typically a few to a few hundred nanometres (millionths of a millimetre). That is also the scale on which physicists, chemists and engineers are now designing ultra-small devices, which is why synthetic biology overlaps with nanotechnology: each can supply components and ideas for the other. This convergence, and the resulting prospect of "cyborg cells," was the subject of our Arctic workshop. It was funded by a philanthropic foundation in California and by research institutes in Delft and Cornell, all established with the backing of physicist and businessman Fred Kavli. Partly as a ploy to ensure that those who attended would stay put, the organisers, Paul McEuen at Cornell and Delft's Cees Dekker, decided to hold it in the town of Ilulissat in Greenland.

There are some obvious reasons why conferences aren't often held in Greenland. It is not an easy place to get to: until May, the only flights came out of Copenhagen, although the Americans attending this workshop took advantage of a newly opened route from Baltimore. Outside the capital of Nuuk, there is nowhere capable of hosting a meeting of any significant size, and in any event Greenlandic travel is not for people on a tight schedule. The introductory session, scheduled for Monday evening, had to be abandoned when the Americans got stranded by fog after landing at Kangerlussuaq (population: 600). Green-land, we are told, is the land of waiting: flight cancellations are routine.

Ilulissat has a population of about 5,000, making it the third biggest town in Greenland. It is unapologetically functional, a perpetual building site perched awkwardly on rocks. There has been no attempt to prettify its gas tanks and container crates—lifelines shipped at much expense from Denmark. It was overcast when we arrived but it felt like we were on the edge of everything. The grey haze where the sea vanishes has a bizarre and frightening quality, a sort of pitiless infinity.

Mindful of the event's resonances, I had packed a copy of Frankenstein, and I could see what Mary Shelley's Arctic explorer Captain Walton meant by being "surrounded by mountains of ice which admit of no escape"—especially when the weather closes in. But when, after an extravagant dinner of juicy scallops, smoked whale and roast musk ox, we went for a walk, the absence of even a hint of dusk was invigorating. Our scramble over ancient granite and through mossy bogs was punctuated by the distant cracks of calving ice.

The fog was back on Tuesday morning, coming and going. They say the weather here can change every 15 minutes. We were tense at breakfast, fearing further delay to our tight schedule. The flight had left Kangerlussuaq, we were told, but would it get in? We rejoiced to hear the aircraft whining low over the hotel. Now we could get down to business.

When, some time soon, bacteria are created with DNA made in a laboratory rather than in some parent cell, like every other cell since the dawn of life, it will be one of the most extraordinary developments of modern science. Yet despite the Faustian overtones, this fresh beginning for life on earth will initially cause few ripples outside the science press.

Why do I predict this muted response? Partly because scientists are unlikely to say that they have "made life"—they will accept, I hope, that such a claim is largely meaningless. But also because, outside science, few people give a damn about bacteria. Philosophers and theologians, ever ready to pronounce on the spiritual status of a single human cell, have no yardstick for evaluating the implications of a synthetic bacterium—and show no signs of acquiring one. The first break in a great chain of being that led from pools of primordial slime to Gordon Brown will prompt little more than the question of whether this stuff is safe.

That is a reasonable, but nonetheless parochial, question. Synthetic life is a technology with the potential to raise, and perhaps answer, issues of genuine philosophical standing. Until the early 19th century, many natural philosophers thought that living matter was different in kind from the fabric of the inorganic world. This so-called "vitalism" was slowly eroded by discoveries in organic chemistry and microbiology. But the idea has by no means disappeared from everyday culture: it can be discerned in the common view of life's sanctity and in the belief in a "divine spark" that somehow transforms a fertilised human egg into a human being. George W Bush's recent second veto of a bill lifting restrictions on US federal funding of stem cell research, on the grounds that it would permit the destruction of "human life," has echoes of this sort of modern-day vitalism.

article body image

But anyone who researches the mechanics of life—who grapples with the question of how a bunch of molecules co-ordinate themselves into an entity that can reproduce, extract energy from its environment, and evolve—is forced to recognise that there is a hazy boundary between the living and the non-living. At this level, the question "What is life?," which has preoccupied scientists from JBS Haldane to Erwin Schrödinger, ceases to have any meaning. Ticking boxes on a supposed list of criteria for "life" is at best an arbitrary process that exposes our inability to think outside the terms of reference we already know. In the end, the argument becomes circular: life consists in all those characteristics we can identify in things we call living. One of the exciting prospects for synthetic biology is that it might permit the exploration of these boundaries, for example by generating "living cells" with non-natural DNA or with some of the protein machinery replaced by designed non-carbon-based devices made with nanotechnology. In any event, the workshop attendees decided rather quickly that the "What is life?" question was a blind alley. They were keen to get on to the issue of how to redesign it.

Of all the smart things one might think of doing with tailor-made microbes, two leap out as urgent: energy and medicine. There is no better location than Greenland for focusing the mind on the need for alternatives to fossil fuels. The north polar region has experienced more warming than most other places on the planet, and it shows. In our village of Ilimanaq, where we were ferried by fishing boat through a frigid sea of blue-green icebergs, we were told by a local schoolteacher that the village hunters find the glaciers more distant every year, and that the thaw water that gushes between the houses each spring has been coming ever earlier. Later, we hiked along the side of the spectacular Jakobshavn glacier, which has retreated many kilometres since the late 19th century. Our guide, a young woman named Vilhelmina, said that the ice floe in the fjord, which produces 10 per cent of all of Greenland's icebergs, had been much thicker when she was a child.

What can designer cells do about this? One of the prime ambitions is to create microbes that will convert plant matter into chemical substances that can serve as fuels. In essence, this is a kind of brewing. Using brewers' yeast to turn grapes into ethanol isn't yet a cost-effective way of making a fuel you can burn in your car, but it might become so if the conversion can be made more efficient. An engine that burns ethanol, or some other biofuel derived from vegetation, still produces carbon dioxide. But this is reabsorbed from the atmosphere when the feed crop is regrown the next season. So you don't get the steady accumulation of atmospheric CO2 that comes from burning fossil fuels. The viability of biofuels is fiercely debated—providing transportation fuel for a large industrialised nation would require vast tracts of land to be given over to fuel crops, and the energy costs of getting the fuels to the pumps are troubling. But the balance sheet can only be improved by making the conversion of plant mass to fuel as efficient as can be. The trouble is that plants resist being broken down—some of the fibrous material is hard for organisms to digest. Yeast or bacteria with fibre-busting genetic machinery borrowed from termites could make better use of the raw material. Steven Chu, a Nobel laureate physicist from the Lawrence Berkeley National Laboratory in California who is immersed in the prickly politics of new energy technologies, claimed that hardy and quick-growing switchgrass could satisfy half of US fuel needs if only it was easier to decompose.

A species of bacterium of the Clostridium genus contains a multiprotein machine that degrades cellulose (a key component of plant cell walls) into glucose. Scientists at the J Craig Venter Institute in Rockville, Maryland, set up by genomics pioneer Craig Venter, want to engineer a species of cellulose-digesting Clostridium that instead generates alcohols like ethanol. John Glass, from the institute, was at the Greenland meeting—a last-minute stand-in for Venter himself, laid low by illness in Costa Rica. Glass suggested that there was no reason to restrict ourselves to ethanol: perhaps we can invent new enzymes that make, say, high-octane hydrocarbons instead.

Giving organisms a suite of genes that allow them to churn out chemical compounds they don't make naturally is a familiar objective in biotechnology, where it is known as metabolic engineering. But with the exception of some relatively simple products such as insulin, cells have not been much exploited as "living factories" for pharmaceuticals. Jay Keasling at Berkeley has spent several years tooling up yeast and bacteria to produce the anti-malarial drug artemisinin, currently extracted at great expense from an Asian shrub. This is one of the most effective drugs on the market, but its prohibitive cost means it has virtually no impact on the formidable death toll from malaria (more than 1m a year worldwide).

Making cells that produce artemisinin means giving them over 40 new genes. These have to be synchronised to avoid bottlenecks in the molecular assembly line, which could cause accumulation of substances toxic to the cells. That means the new genes can't simply be slotted into the cells' genomes: Keasling has to grapple with the genetic "circuit diagram" that determines when and how genes are switched on and off. This is one of the issues that distinguishes synthetic biology from traditional genetic engineering: it involves intervening with the genome, the cell's operating system, as if rewiring a complex electrical circuit. There are feedback loops, amplifiers, switches and control circuits that ensure genes operate in synchrony. But the understanding of this systems-scale cell logic is still rudimentary.

Yet Keasling's work shows that the problem isn't intractable. He has now created forms of yeast that make large amounts of artemisinic acid—just a few simple chemical steps away from artemisinin itself—and pump it out of the cells for easy extraction and purification. He hopes to see an industrial process up and running by 2009, making the drug at the affordable cost of ten cents per gram.

The artemisinin story is a reminder of how the pharmaceutical industry still depends on so-called "natural" products—chemical compounds made in nature. Despite grand talk of "drug design" and the promise of new technologies such as antisense therapy, in which problematic genes are "silenced" by DNA-like molecules, most drugs now in use are natural products or compounds derived from them, such as aspirin. Yet the drug pipeline is drying up: fewer are approved now than 30 years ago. One of the hurdles many drug candidates fail to clear is that, even if they work well for most people, there may be small subsets of the population whose genetic make-up will make them prone to nasty, even fatal, side-effects. At the moment, such would-be drugs are binned. But they might be rescued by a quick, easy and reliable means of spotting individuals with the "wrong" genes. These genes can potentially be identified from the proteins they encode, present in blood samples, say. But that's a classic needle-in-haystack affair. Nanotechnology promises to deliver sensors that detect biochemicals highly selectively at very low levels. Scott Fraser of the California Institute of Technology talked about sensors that work like miniature tuning forks set ringing by laser beams: if just a few of the right molecules stick to their surface, the ring changes pitch. This, he said, was what the much-hyped "personalised medicine" should really entail. Decoding individual genomes so as to prescribe drugs tailor-made to our personal genetic constitution is impractical economically and perhaps even scientifically. But screening for common genetic conditions that make good drugs do bad things would at least reduce the crippling rate of attrition in drug discovery.

Things were warming up. Throwing a hugely diverse bunch of scientists together in an isolated outpost on the edge of the world was a risky experiment, not least because they might lack a common technical language and culture. But despite the awesome specialist knowledge in attendance, no one was an expert on how to "make life." "We are all bozos on this bus," was how biophysicist Bob Austin of Princeton put it. By Wednesday night, everyone had begun to shed the burden of expertise and let their imaginations roam. Sadly, the Americans were leaving for Baltimore after lunch the following day.

There are two ways of approaching synthetic biology. One works from the bottom up: starting with chemical reagents taken from the lab store, can you make something that behaves even a little like a living cell? The other is top-down: you take the parts and design principles from natural organisms but then put them together differently. Many regard the first approach as far too difficult, a little like trying to make a computer from wire, solder and a high-school understanding of electronics. All the same, the bottom-up synthesis of living things from crude ingredients must be possible, since it happened of its own accord around 3.8bn years ago—an event for which the earliest evidence comes from rocks just south of Ilulissat. Attempting to generate lifelike systems chemically, however ambitious, might at least address some of the big questions about that event. Which environments are most fecund? What building blocks do you need? Where does the organisation come from? Which comes first: replication or compartmentalisation? Were the earliest cells "garbage bags," as veteran physicist Freeman Dyson put it, containing an uncoordinated mix of stuff, or were they pared-down entities with nothing superfluous?

The top-down approach tends to the view that simplification is a good strategy for designer cells, regardless of whether early terrestrial life showed such economy. We still have little grasp of the rules that govern networks of gene interactions even in rather "primitive" bacteria, but the best hope of untangling them, and thus of exercising rational design, comes from identifying the "minimal cell": the smallest subset of genes needed to sustain a viable organism. The notion of "minimal" is treacherous here, because it depends on context. If you provide a cell with all the amino acids from which proteins are made, say, it no longer needs enzymes for synthesising them. But then it wouldn't live long in the wild. The same is true of genes for coping with temperature changes or attacks by other organisms. Evolution depends on stresses like these which weed the strong from the weak. That's why so much of an organism's genome is devoted to coping with them.

If your designer bacteria that chew up grass and generate fuels are bred in carefully regulated isolation, that may not matter. But it might be best to make them at least heat-tolerant, so that they thrive while contaminating microbes get boiled. In any event, scientists at the Venter Institute are so confident they have identified a set of "minimal genes" that they have applied for a patent. (I'm told it's unlikely to be granted.) This set comprises 381 genes in the genome of a bacterium called Mycoplasma genitalium, a parasite of the human genital tract with a remarkably small genome of just 485 protein-coding genes. If Venter's team is right, the stripped-down M genitalium genome will be the chassis on which new bacteria will be designed and built more or less from scratch.

How do you build a bacterium with a redesigned genome? DNA—the genetic material—can now be made in long lengths in the lab using chemical methods, with a precisely specified genetic message defined by the sequence of its four chemical building blocks (bases) along the chain. Around 40 companies now supply tailored DNA at around $1 per base—a cost that has been halved each year for the past decade or more. Making a chain as big as an entire bacterial genome remains challenging, but it should be possible to build it in segments and join them together.

Yet a genome is no more an organism than a CD of a computer's operating system is a laptop. How do you reincarnate a naked genome as a living bacterium? In the meeting's most dramatic revelation, John Glass explained how. He and his colleagues have performed the first whole-genome transplant: removing the genome of one species of Mycoplasma and inserting it into another. The recipient cells can be "booted up" with the new operating system, and begin behaving just like the donor cells. It's a little like creating an acorn that grows not into an oak but a willow. This surely presages the insertion of a wholly synthetic genome: the Venter Institute scientists will try to use their technique to make cells containing their putative minimal Mycoplasma genome and see if they survive.

One might imagine that using nature's tricks to make stuff has a green tinge to it, but there's no telling where the ethical arguments might end up: Angela Belcher of the Massachusetts Institute of Technology, who adapts viruses to make nanotechnological machines, said she has been denounced for "cruelty to viruses." It is proper, however, that the imminence of the first human-made organisms should give us pause. Paul McEuen, a specialist in nanoelectronics, confessed that his sleep was disturbed more by what he had heard than by the midnight sun. The risk of mistakes—of synthetic microbes escaping into the environment or mutating uncontrollably—is bad enough. Yet one can imagine giving an artificial genome a variety of safety features—for example, making the organism dependent on a nutrient that doesn't exist in nature, or giving it a "fail-fast" genetic code different from the one natural DNA uses, such that every mutation proves fatal and evolution is consequently proscribed.

But the thought of bacteria and viruses being deliberately concocted by people wanting to wreak havoc is alarming. Drew Endy of MIT pointed out that making the genome of the Ebola virus now costs about as much as a new Volkswagen. In three years it might cost the price of a laptop, and in five to seven years, that of an iPod.

Yet shelving synthetic biology on the grounds that it is too dangerous simply isn't an option. Now that the technologies exist to reinvent what an organism is, the real question is how we can best ensure that they are used safely and with good intent. Of course, it would be hard to identify a useful technology that has not been militarised (many began in defence labs), and fledgling synthetic biologists are undoubtedly already at work in military establishments. Endy admitted that he knew no more than anyone else about what transpires in the US National Biodefense Analysis and Countermeasures Center at Fort Detrick in Maryland, even to the extent of whether its research is solely defensive. Meanwhile, Make magazine, the bible of American home inventors, has already shown its readers how to do "backyard biology," and the idea that inventing your own biological virus might become as easy and attractive to disaffected technophiles as devising computer viruses is obviously disturbing. While it makes sense to insist that DNA-synthesis companies vet their orders, and perhaps that access to pathogen genome data be restricted, that can't close every leak.

Such broader questions benefit from historical perspective, and few scientists can provide as much of it as Freeman Dyson. One of the most influential and politically aware of the post-Manhattan project physicists, he is now 83 but still game for a freezing midnight ride through icebergs. For all its genius, his generation failed to forsee the technological future, he said: "We totally missed all the important things." He recalled how his former Princeton colleague John von Neumann, one of the founders of computer science, estimated that the US would only ever need 18 computers. Dyson himself confessed to once trying to persuade Francis Crick against moving into biology.

It was too early to leave, but time had run out. Another day, we suspected, and the discussion would really have kicked into gear. But if you don't get out of this place on Friday, you're there for the weekend. The organisers and I have drawn up a declaration—the "Ilulissat statement" (tinyurl.com/35nkgn)—to which all the attendees subscribe, proposing in broad outline how we think the field should develop. It's a useful exercise, yet I can't help thinking of Dyson's words. Far from being disheartening, however, I find them rather exciting.